Re: [PATCH 08/13] target/riscv/vector_helper.c: use vlenb in HELPER(vsetvl)

[Richard Henderson]

On 1/13/24 08:38, Daniel Henrique Barboza wrote:
> Use the new 'vlenb' CPU config to validate fractional LMUL. The original
> comparison is done with 'vlen' and 'sew', both in bits. Use sew/8, or
> sew in bytes, to do a direct comparison with vlenb.
>
> Signed-off-by: Daniel Henrique Barboza <dbarboza@ventanamicro.com>
> ---
>   target/riscv/vector_helper.c | 9 ++++++---
>   1 file changed, 6 insertions(+), 3 deletions(-)
>
> diff --git a/target/riscv/vector_helper.c b/target/riscv/vector_helper.c
> index cb944229b0..0c1a485d1e 100644
> --- a/target/riscv/vector_helper.c
> +++ b/target/riscv/vector_helper.c
> @@ -45,9 +45,12 @@ target_ulong HELPER(vsetvl)(CPURISCVState *env, target_ulong 
> s1,
>                                               xlen - 1 - 
> R_VTYPE_RESERVED_SHIFT);
>   
>       if (lmul & 4) {
> -        /* Fractional LMUL - check LMUL * VLEN >= SEW */
> -        if (lmul == 4 ||
> -            cpu->cfg.vlen >> (8 - lmul) < sew) {
> +        /*
> +         * Fractional LMUL - check LMUL * VLEN >= SEW, or
> +         * vlen >> (8 - lmul) < sew. We have 'vlenb', so
> +         * compare it with sew in bytes.
> +         */
> +        if (lmul == 4 || cpu->cfg.vlenb >> (8 - lmul) < (sew << 3)) {

You're shifting sew the wrong direction.  But better to simply adjust the vlenb 
shift:

- cpu->cfg.vlen >> (8 - lmul) < sew
+ cpu->cfg.vlenb >> (8 - 3 - lmul) < sew


r~